Diffusion tensor characteristics of gyrencephaly using high resolution diffusion MRI in vivo at 7T

Michiel Kleinnijenhuis, Tim van Mourik, David G. Norris, Dirk J. Ruiter, Anne-Marie van Cappellen van Walsum, Markus Barth

Gyrification of the human cerebral cortex allows for the surface expansion that accommodates many more cortical neurons in comparison to other mammals. For neuroimaging, however, it forms a feature that complicates analysis. For example, it has long been established that cortical layers do not occupy the same depth in gyri and sulci. Recently, in vivo diffusion imaging has provided insights into the fibre architecture of the cortex, usually showing radial tensor orientations. This makes it relevant to investigate whether cortical diffusion tensor metrics depend on the gyral pattern. High-resolution (1mm isotropic) diffusion weighted MRI of the medial wall of the hemispheres was performed at 7T. Diffusion data was resampled to surfaces in the cortex and underlying white matter, where the cortical surfaces obeyed the equivolume principle for cortical laminae over the cortical curvature. Diffusion tensor metrics were averaged over bins of curvature to obtain maps of characteristic patterns in the gyrus. Diffusivity, anisotropy and radiality varied with curvature. Radiality was maximal in intermediate layers of the cortex next to the crown of the gyrus, not in white matter or on the crown. In the fundus, the deep cortical layers had tangential tensor orientations. In the white matter, tensor orientation changed from radial on the crown to tangential under the banks and fundus. White matter anisotropy gradually increased from the crown to the fundus. The characteristic pattern in the gyrus demonstrated here is in accordance with ex vivo diffusion MR microscopy and histological studies. The results indicate the necessity of taking into account the gyral pattern when cortical diffusion data is analysed. Additionally, the data suggest a confound for tractography approaches when reaching the gyrus, resulting in a possible bias towards the gyral crown. The implications for mechanisms that could drive cortical folding are discussed.